EPANET and Tanks

[JonathanEngr]

I have a water system model that I'm connecting to another "dummy" system. The system that I'm connecting it to is significantly lower in elevation (the new system and tank are about 200 feet lower than the current system). The new system consists of a tank that is filled with my existing system, and the water flows out of this tank into the "dummy" system that is simply a node with a demand curve. Basically, I'm trying to see how this new system will affect pressures within the current system.

However... when I run an analysis, the model seems to allow the new tank to overflow. The tank is 70,000 gallons... it is 20' in diameter and 30' tall, and the initial setting for depth is 20 feet (2/3rds full). It is connected to the existing system via a 10-inch water line. From the very start, the flow into the tank is roughly 1,000 gpm, and it stays that way throughout the full 24-hour analysis period despite the fact that the only line leaving the tank goes to my demand node. It is correctly reflecting the usage, which ranged from 220 to 660 gpm. No water is flowing in the other direction due to elevation (and the flow direction arrows confirm this). I thought tanks in EPANET are considered closed systems... i.e., they don't overflow--they just close. However, when observing the tank data, there is no indicator for "open" or "closed" like other objects in EPANET.

Where is my water going? It's wreaking havoc with my model... any thoughts?

 

[BigInch]

I don't use EPAnet, but most hydraulic simulators model "tanks" to overflow, if you don't put a level controlled open/close valve at the inlet. As I understand, EPAnet is a finite state, step to step, modeller, so if there is more inlet flow than outlet flow and the tank becomes full, it wouldn't be able to handle the transient calculations needed to constantly back up the flow into the inlet piping, increasing the pressure and slowing the flow in the inlet piping system.

Independent events are seldomly independent.

 

[JonathanEngr]

I wondered that, myself, but after searching EPANET and tanks on the web, it appears that it uses a closed system to keep that from happening. Here's link to a discussion about this topic:

http://communities.bentley.com/products/hydraulics___hydrology/f/5925/p/7430/11435.aspx

The quote from that discussion states:

-------------------------------
That is how the EPANet based Calculation engine works, as there is no overflow from Tanks, they are treated as a "closed" system. When a tank reaches 100% this triggers the inlet pipes to close, and the calculation results returned are as if the inlet pipes are closed (if you click on them on them during the right timestep it will say Calculated Status = Closed), and the tank will not come back into play in the model until the simulation hydraulics allow water to drain out of the tank.
-------------------------------

However, on my tank info there is no "status" for open or closed. Perhaps this was an earlier version? I did note that back in 2010 they announced EPANET 3.0, which (to my knowledge) hasn't been released. They were going to include altitude valves in this release I'm assuming for this purpose. I've even tried dropping in PRV's and PSV's to see if I could get them to shut off the flow, but no luck yet. It's quite frustrating, as the model is an excellent one that has been calibrated and works beautifully. But trying to tie another "system" onto it has been a nightmare!

Do you know if there is any other software out there at a more budget-minded price than WaterCAD? I really like WaterCAD, but to get the software with enough nodes for my use it is prohibitively expensive since it's not something I would use very often.

 

[BigInch]

"inlet pipes are closed" I think that's what I said, if that can mean there is a (in that case a virtual) level control valve on the pipe that actually "closes the pipe".

I'm afraid I only know the really expensive ones.
Have you looked at AFT Flow line? Sorry I don't know the cost of that one?

Independent events are seldomly independent.